In Chapter 15, I fit a Poisson regression of number of interlocks on assets, nation of control, and sector for Ornstein’s Canadian interlocking-directorate data. The results from this regression are...

In Chapter 15, I fit a Poisson regression of number of interlocks on assets, nation of control, and sector for Ornstein’s Canadian interlocking-directorate data. The results from this regression are given in Table 15.3 (page 428). Influential-data diagnostics (see, e.g., Figure 15.7 on page 456) suggest that the first observation in the data set is quite influential; in particular, the coefficient of assets changes considerably when the first observation is removed. Perform a robust Poisson regression for this model. How do the results compare to removing the first observation from the data set? (Recall, however, that the influence of the first observation depends on unmodeled nonlinearity in the relationship between interlocks and assets—a problem that I ultimately addressed in Chapter 15 by log-transforming assets.)